Automated detection of trinucleotide repeats in fragile X syndrome

Molecular Diagnosis Vol. 2 No. 4 1997

Automated Detection of Trinucleotide Repeats in Fragile X Syndrome HASNAH

HAMDAN,*

JOHN A. TYNAN, t RAYMOND

A. F E N W I C K , t

JORGE A. LEON* San Juan Capistrano, California

Background: The conventional method for diagnosis of fragile X syndrome has been amplification of the trinucleotide repeat region of the FMR-1 gene by polymerase chain reaction (PCR) and Southern blot analysis to detect full expansion and hypermethylation. "Stuttering" resulting from incomplete amplification is still observed in the PCR products despite the use of reagents that reduce the secondary structure of the GC-rich template. In addition, PCR products can be detected by autoradiography only after 1 to 2 days of exposure. By combination of a recently reported amplification protocol with fluorescence detection of PCR products in an automated DNA sequencer, the PCR protocol for amplification of trinucleotide repeats was simplified. This modified protocol is highly reproducible, more accurate, and less costly than the conventional protocol because of the elimination of radioisotopes from the PCR. Methods and Results: PCRs were conducted with betaine and Pju DNA polymcrasc. This improved PCR protocol allowed immediate detection of PCR products in agarose gels containing ethidium bromide. Stuttering was completely eliminated and fragments of up to 1 kb (~250 repeats) wcrc visible in agarose gels. PCR products were automatically detected by laser fluorescence in an automated DNA sequencer by inclusion of a fluorescently-labeled primer in the PCR reaction. A short electrophoresis run of 100 minutes in denaturing acrylamide gels was sufficient to give high resolution of fragments with higher accuracy and sensitivity than conventional detection by autoradiography. Conclusions: A simple, nonradioactive protocol that is more rapid and less expensive than the conventional PCR protocol for the detection of trinucleotide repeats has been developed. By use of this detection protocol, fragment sizes containing up to 100 repeats could be detected, alleles differing by one trinucleotide repeat were clearly resolved, and heterogeneous repeat patterns such as those present in mosaics could be discriminated. This protocol has been adapted to the amplification and detection of at least two other classes of trinucleotide repeats [(CAG)~ and (CTG)n], suggesting that it may be a universal protocol for PCR amplification and detection of trinucleotide repeats. Key words: genetics, polymerase chain reaction, fluorescence, betaine.

From the Departments of *Biotechnology Research and Development and "Molecular Biology, Quest Diagnostics at Nichols" lust# tute, San Juan Cupistrano, California. Supported by a grant from Pharmacia Biotech.

Reprint requests: Hasnah Hamdan, PhD, Department of Biotechnology Research and Development, Quest Diagnostics at Nichols institute, 33608 Ortega Highway, San Juan Capistrano, CA 92690. Copyright © 1997 by Churchill Livingstone ®

1084-8592/97/0204-000458.00/0

259

260

Molecular Diagnosis Vol. 2 No. 4 December 1997

Fragile X (FRAX) syndrome is an X-linked genetic disease characterized by the expansion of the trinucleotide repeat CGG within the 5' untranslated region (UTR) of the F R A X mental retardation 1 (FMR-1) gene [1-3]. The number of CGG repeats in normal, premutation, and full mutation alleles range from 5 to 50, 51 to 200, and greater than 200 repeats, respectively [4]. The expansion in CGG repeats and DNA methylation disrupt the function of the FMR-1 gene on the X chromosome [5]. Patients carrying the full FRAX mutation exhibit severe mental retardation and abnormal growth development [6]. The conventional method for diagnosis of FRAX syndrome is amplification of the trinucleotide repeats by polymerase chain reaction (PCR) and Southern blot analysis of genomic DNA to detect expanded repeats and DNA hypermethylation [7]. Conventional PCR protocols for FRAX include the use of 7-deaza-2'deoxyguanosine triphosphate and dimethyl sulfoxide to reduce secondary structure in the DNA template, which allows for better amplification; however, PCR products from the conventional protocol are not visible in agarose gels stained with ethidium bromide. Thus, 32P-labeled deoxynucleotide triphosphates have to be incorporated into the PCR so that PCR products can be detected by autoradiography, which requires exposure of 1 to 4 days. Multiple bands or "stuttering" also is often observed with the conventional PCR protocol due to pausing of the Taq DNA polymerase [7,8]. At least two reports have recently been published [9,10] that describe improved amplification of triplet repeats. These protocols include the use of the osmoprotectant betaine [11] to reduce secondary structure in the template [12] and the proofreading enzyme Pfu DNA polymerase in combination with Taq DNA polymerase to increase the fidelity of amplification. PCR products obtained with this protocol are easily visualized in agarose gels stained with ethidium bromide and appear as distinct bands with virtually no stuttering [9]; however, the number of repeats cannot be accurately determined by electrophoresis in agarose gels. By including a fluorescently-labeled primer in the PCR, we were able to use laser fluorescence to automatically detect and accurately size F R A X PCR products by electrophoresis in sequencing gels and to eliminate the use of radioisotopes in the PCR. Here we report the automated detection and siz-

ing of improved F R A X PCR products from clinical samples.

Materials and Methods Genomic DNA Extraction Total DNA was extracted from 5 mL of whole blood in acid-citrate-dextrose or ethylenediaminetettraacetic acid by proteinase K digestion [13] followed by phenol:chloroform as described previously [14]. D N A was diluted in 10 mM Tris-HC1, i mM ethylenediaminetettraacetic acid, pH 8.0 (TE) to a final concentration of 20 ng/txL for PCR.

PCR Amplification A single-tube PCR was developed for amplification of triplet repeats in the FMR-1 and androgen receptor (AR) genes. The PCR protocol was conducted exactly as described previously [9] except for the following modifications: (1) the reaction was scaled down to 10 txL; (2) the F R A X primers were reduced to a final concentration of 1.2 ng/txL each; (3) AR primers (exon 1, CAG repeats) were added at a final concentration of 1.2 ng/txL each and one primer in each set was fluorescently labeled with Cy-5 (Operon Technologies; Alameda, CA). We routinely prepare our own PCR kits containing all reagents except for Pfu DNA polymerase (Stratagene; La Jolla, CA), AmpliTaq DNA polymerase (Perkin-Elmer; Foster City, CA), and genomic DNA as follows: 500 txL 10X Pfu buffer (Stratagene); 10 txL each of 100 mM deoxyadenosine, deoxycytidine, deoxyguanosine, and deoxythymidine triphosphates; 60 txL each of FX-5C, FX-3F, AR-F, AND AR-R primers (stock concentration of 100 ng/p~L each); 2500 ~L of freshly prepared 5 M betaine monohydrate (Sigma); 1000 p~L dimethyl sulfoxide (Pharmacia Biotech); and 520 IxL sterile water for a final volume of 3900 txL. Polymerase chain reaction kits were dispensed in 200-txL aliquots and stored at -70°C. Kits were stable for at least 6 months under these conditions. After addition of the respective enzymes (Pfu DNA polymerase and AmpliTaq DNA polymerase to final concentrations of 0.04 and 0.02 U/txl, respectively), 8 ~L of PCR mix was dispensed into respective 200-txL PCR tubes. Two microliters of genomic DNA (20 ng/lxL) was added

Automated

and the tubes were sealed with a MicroAmp Full Plate cover (Perkin-Elmer), placed in a thermal cycler (Perkin-Elmer 9600), and amplified as follows: 95°C, 6 minutes followed by 30 cycles of 95°C, 1 minute; 60°C, 2 minutes; 75°C, 5 minutes; and 1 cycle of 75°C for 13 minutes as previously described [9]. The sequences of the PCR primers are as follows: F R A X - - s e n s e (FX-5C), 5'-(Cy5) G e T CAG CTC CGT TTC G G T TTC ACT TCC G G T 3'; antisense (FX-3F), 5' AGC CCC GCA CTT CCA CCA CCA GCT CCT CCA 3' [7]: A R - - s e n s e (AR-F), 5'-(Cy5) ACC A G O TAG CCT GTG G G G CCT CTA CGA CGA T G G GC 3'; antisense (AR-R), 5' CCA GAG CGT GCG C G A A G T GAT CCA G A A CCC GG 3' [9]. All PCR primers were purified by high-performance liquid chromatography (Operon Technologies), reconstituted in sterile water to a stock concentration of 100 ng/IxL, and stored at -20°C. Cy5-1abeled primers were stored at -20°C wrapped in aluminum foil to prevent excessive exposure to light. Detection of PCR Products

Verify Amplification in Agarose Gels. Five microliters of PCR products was analyzed in 1.5% agarose gels (FMC BioProducts) containing ethidium bromide (20 ng/IxL). Electrophoresis was conducted at 200 V for 30 minutes in IX Tris-borateethylenediaminetettraacetic acid buffer [14]. Confirm Size in High-Resolution Polyacrylamide Gels. Polyacrylamide gels were prepared as follows: Short plate cassettes designed for the ALFExpress DNA sequencer (Pharmacia Biotech) were used with 0.3-mm quartz spacers (resulting gel size is approximately 14 × 29 cm). After addition of ammonium persulfate as per manufacturer's instructions, approximately 30 mL of ReadyMix gel (6% acrylamide, 7 M urea, 100 mM Tris-borate, pH 8.3, 1 mM disodium ethylenediaminetettraacetic acid, and 3 mM TEMED, Pharmacia Biotech) was injected into the cassette. The gel was allowed to polymerize at room temperature for at least 2 hours before electrophoresis. The wells were rinsed twice with 0.6X Tris-borate-ethylenediaminetettraacetic acid buffer before the samples were loaded. Samples to be loaded in acrylamide gels were prepared as follows: 5 IxL of formamide-loading buffer (Pharmacia Biotech)

Detectionof Trinucleotide Repeats

•

H a m d a n et al.

261

containing Cy5-1abeled sizer molecules (100 and 300 bp to serve as internal standards; Pharmacia Biotech) at a concentration of 1 fmol/p.L was added to 5 txL of each PCR product. Samples and external standards were heated at 95°C for 5 minutes in a thermal cycler and rapidly cooled on ice, and then 7.5 p~L of a denatured PCR product was loaded per well. A similar volume of Cy5 Sizer 50-500 bp (Pharmacia Biotech) diluted to a final concentration of 2 fmol/txL in 10 mM Tris-HC1, 1 mM ethylenediaminetettraacetic acid, pH 8.0, and formamide loading buffer was loaded in the middle lane to serve as the external standard. Electrophoresis in polyacrylamide gels was conducted in an ALFExpress sequencer under the following conditions: 0.6X Tris-borate-ethylenediaminetettraacetic acid buffer, 1500 V, 60 mA, 25 W, 55°C, sampling interval of 2 seconds, and run time of 100 minutes. Solid-Phase Sequencing of PCR Products To validate the amplification of triplet repeats using the modified protocol, F R A X PCR products from three male samples were sequenced using a solid-phase sequencing kit (Pharmacia Biotech) and analyzed in the ALFExpress sequencer. PCRs to generate products to be sequenced were conducted with a biotinylated sense F R A X primer (FX-5C) and a nonbiotinylated antisense F R A X primer (FX-3F). Forty microliters of PCR products was then sequenced as per manufacturer's instructions using the following sequencing primer: 5'-(Cy5) CTT CTC TTC A G e CCT G e T A G T 3' (4 pmol per reaction). Method Comparison To compare the efficiency of our modified protocol with that of the conventional P e R protocol we analyzed 20 samples (10 males, 10 females) by both methods. The size of the trinucleotide repeat expansion in these samples had previously been determined at Nichols Institute using the conventional protocol described by Fu et al. [7] and ranges from 19 to 79 repeats and from 23 to 99 repeats for the samples from males and females, respectively. In addition, 60 clinical samples (31 males, 29 females) were analyzed to determine the efficiency, accuracy, and sensitivity of our automated detection system for determining the number of triplet

262

Molecular Diagnosis Vol. 2 No. 4 December 1997

the number of repeats was determined automatically using the formula [size(bp) - 221]/3 [7].

repeats in FRAX. The number of repeats in these samples was determined simultaneously by both the conventional and modified protocols. Detection of PCR products was done by autoradiography for the conventional protocol and by automated laser fluorescence for the modified protocol. The size of fragments obtained by the conventional protocol was determined visually by referring to M13 size ladders run alongside [7]. The size of fragments obtained by the modified protocol was automatically determined using a fragment analysis program (Fragment Manager Version 1.0 from Pharmacia Biotech). Once the size of fragments had been determined, the information was entered into a spreadsheet (Microsoft Excel), and

1

2

3

4

5

Results Detection of PCR Products

Analysis in Agarose Gels. All samples amplified strongly in PCRs for FRAX. PCR products from males appear as distinct, single bands with very little stuttering or heteroduplexing (Fig. 1). Both F R A X alleles in females amplified uniformly and PCR products were clearly visible as two distinct bands in agarose gels; however, some hetero-

6

7

8

9

10

11

12

500 b

250 b

Fig. 1. Detection of single, discrete PCR products from amplification of trinucleotide repeats in fragile X alleles of samples from males. In this experiment, PCR was conducted only with fragile X primers at a final concentration of 2.4 ng/IxL. Five microliters of PCR product was loaded per lane and electrophoresed in 1.5% agarose gels containing ethidium bromide (20 ng/ixL). Lane 1: 50bp DNA ladder (Pharmacia Biotech; Piscataway, NJ); lanes 2-11: samples with 79, 62, 52, 44, 40, 35, 31, 29, 23, and 19 repeats, respectively; lane 12: negative control (no template DNA in PCR reaction).

Automated DetectionofTrinucleotideRepeats

1

2

3

4

5

6

7

8

9

10

11

•

Hamdanetal.

263

12

800bp 250 bp

AR

Fig. 2. Duplex amplification of trinucleotide repeats in fragile X and androgcn receptor alleles of samplcs from females. PCR was conducted as dcscribcd uudcr Materials and Methods and 5 bLL of P('R product was Ioadcd per hmc and analyzcd in 1.5% agarose gels containing cthidium bromide (20 ng/FL). Lane 1: 50-bp DNA ladder (Pharmacia Biotcch); hines 2 I l: samples with ~200/30, 99/29, 99/30, 85/30, 78/29, 67/29, 58/29, 50/30, 43/30, and 32/23 repeals, respectively: lane 12: negativc control (no tcmplate DNA added to PCR reaction).'l]~c position of androgen receptor PCR products is indicated by the right arrow.

duplexing was observed in the larger allele, and PCR bands from amplification of the large alleles (>80) were fainter when compared with the bands for the normal allele (Fig. 2).

Analysis in High-Resolution Polyacrylamide Gels. P C R products analyzed in thin sequencing gels were detected as distinct, single peaks with no extraneous bands. There was minimal background, and heteroduplexing was completely eliminated (Fig. 3). Products of lower molecular weight (<450 bp) were detected as sharp peaks; however, peaks started to broaden with increase in the fragment size (Fig. 3C).

Limits of Detection Using the P C R protocol containing betaine, we were able to amplify alleles of up to 200 repeats yielding products of approximately 1 kb. PCR products of this size were faintly visible in agarose

gels (Fig. 2) but could not be detected by laser fluorescence. Use of a fluorescently labeled primer in the PCR protocol allowed the reliable detection of products up to 500 bp (approximately 100 repeats) (Fig. 3C) in polyacrylamide gels, whereas fragment sizes up to 1 kb can be detected in agarose gels.The upper limit of amplification was approximately the same for males and females; however, visual comparison of bands in agarose gels indicates that expanded alleles from males are more readily amplified than similar-sized alleles from females.

Method Comparison Study The number of repeats determined by the modified protocol was concordant with that of the conventional protocol in only 20% (12/60) of the cases studied. In each of the nonconcordant cases, the number of repeats determined by the modified protocol was 1 or 2 repeats larger than that deter-

264

Molecular Diagnosis Vol. 2 No. 4 December 1997

'I @

<

rn

m--

0

M o--r

X 0.~ m~ ~

~.~.

~

~,

o

~

m~

~ ~.°~

(",1 ~

x ~ x

~

~-.~ ~= C~

J ~

J o'~ o

o

0 0

0.~::

N~'-"

0

Automated Detection of Trinucleotide Repeats

mined by the conventional protocol (Table 1). This discrepancy was resolved by sequencing the PCR products from two male samples (1470 and 1531) in which the number of triplet repeats (20 and 30 repeats) was different from that determined by the conventional protocol. The PCR product from a male sample (1474) in which the number of triplet repeats was in concordance with that determined by the modified protocol was also sequenced. Sequencing showed the presence of 20 and 30 CGG repeats for samples 1470 (Fig. 4) and 1531, respectively; 29 repeats were observed in sample 1474. Sequences from all three samples showed the presence of AGG repeats interspersed after every 10 repeats, as has been reported previously [15]. Thus, we conclude that the modified PCR protocol combined with high-resolution electrophoresis and automated detection of PCR fragments resulted in more accurate sizing of triplet repeats. Using the modified protocol, we detected 30 repeats as the most common allele (35%) in the 31 males studied, followed by 29 (16%), 31 (13%), and 20 (10%) repeats, respectively. The most comm o n allele in the 29 females studied also was 30 repeats (33%), followed closely by 29 repeats (27%). These results agree with previously r e p o r t e d [16] frequencies for the most c o m m o n F R A X allele.

Discrimination of Alleles Our method comparison study highlighted the inability of the conventional protocol to discriminate reliably the second F R A X allele in samples f r o m females when sizes of the two alleles differ by only 1 or 2 repeats. In three of the 29 samples analyzed (samples 1483, 1585, 1589), the second F R A X allele was not differentiated by the conventional protocol, which identified these cases as h o m o z y gous for the F R A X alleles with 29, 28, and 29 repeats, respectively (Table 1). Analysis of P C R products from these three samples in our highresolution acrylamide gel system, however, clearly resolved the two alleles in each case, thus determining t h e m as h e t e r o z y g o u s for the F R A X allele with 29/28 repeats (Fig. 3A).

Duplexing of AR and FRAX PCR Inclusion of the AR primers (exon 1, CAG) in the P C R serves two functions: (1) as a control for the presence of two alleles in female samples hom o z y g o u s for F R A X and (2) as a control for the

•

Hamdan et al.

265

Table 1. Summarized Results of Study Comparing the Number of Triplet Repeats Detected in 60 Clinical Samples Using the Conventional and Automated Fragile X PCR Protocols No. of CGG Repeats Sample No. 1535 1470 1472 1473 1537 1495 1569 1573 1469 1476 1529 5882 1471 1474 1531 1534 1536 1539 540 1572 1574 1577 1583 1532 1568 1571 1576 1477 1494 1581 1589 1590 1582 1580 1544 1547 1487 1548 1586 1530 1488 1585 1481 1497 1578 1533 1478 1480 1483 1486 1588 1589 1549 1570 1475 1538 1546 1482 1575 1517

Conventional 15 19 19 19 22 24 27 28 28 28 28 29 29 29 29 29 29 29 29 29 29 29 29 30 30 30 30 31 39 38 47 34/15 22/19 29/19 23/19 29/19 29/19 37/19 29/22 29/22 39/24 28/28 28/28 29/28 29/28 29/28 29/29 29/29 29/29 29/29 29/29 29/29 30/28 30/29 30/29 31/29 32/29 34/29 43/29 48/48

Automated

Sex

17 20 20 20 23 24 29 28 30 29 29 30 31 29 30 30 29 31 30 30 29 30 30 31 31 32 30 32 38 39 47 35/16 23/20 30/20 24/20 30/20 29/19 39/20 30/22 32/24 39/24 29/28 28/28 30/29 30/29 30/29 31/30 30/30 29/28 29/29 30/30 29/28 31/29 31/29 31/30 32/30 33/30 34/29 44/29 49/49

Male Male Male Male Male Male Male Male Male Male Male Male Male Male Male Male Male Male Male Male Male Male Male Male Male Male Male Male Male Male Male Female Female Female Female Female Female Female Female Female Female Female Female Female Female Female Female Female Female Female Female Female Female Female Female Female Female Female Female Female

266

Molecular Diagnosis Vol. 2 No. 4 December 1997

Bases

95

90

85

80

75

70

65

60

55

50

45

40

35

30

25

20

CGr GCG3CAGC3CC~CGGCGGCGGCGGCGGCGC3~CGGCGGA GGCGGCGGCGGCGGCGGCGGCGC_.-~G~..C~..~TG33CCT C~G3 5'

1 / 2 /3

/ 4 / 5 / 6 / 7 / 8 / 9 /10 / 1 1 / 1 2

/ 1 3 / 1 4 / 1 5 /16 /17 /18 /19 / 2 0

Fig. 4. Confirmation of the accuracy of automated sizing of F R A X alleles. The number of repeats in this male patient (sample 1470) was determined as 19 by the conventional protocol, but was calculated to be 20 repeats by our automated protocol (see Table 1). The PCR product from amplification of the normal F R A X allele from this sample was sequenced as described under Materials and Methods. Sequence data showed the presence of 20 C G G repeats (numbered 1-20), with an A G G repeat interspersed at repeat 11.

presence of D N A template in samples of affected males with fully expanded repeats that are not amplifiable.The peak for the F R A X P C R product in a female homozygous for the F R A X allele appears as a single peak approximately twice the height of the individual A R peaks (Fig. 3B). The A R primers at the concentrations used in our protocol do not seem to interfere with the amplification of the F R A X alleles. We have found that, occasionally, the A R product is of the same size as one of the F R A X products, and thus the two peaks overlap and cannot be differentiated because the detection method used is a one-dye system. This problem is easily resolved by repeating the amplification for F R A X and A R in two individual reactions. In addition to the duplex P C R kits, we also routinely prepare P C R kits containing only F R A X and A R primers to be used for resolving these rare cases.

Discrimination of Mosaics Using the modified protocol, we were able to detect and identify a male who was mosaic for the F R A X allele. A sample from the same patient had been tested earlier using the standard Southern blot analysis and radioactive P C R protocol [7], but the abnormality was missed. The case had been reported as normal with 37 repeats and appeared normal by Southern blot analysis; however, the patient had a reported phenotype consistent with F R A X syndrome. A sample was resubmitted and analyzed using our modified protocol and standard Southern blot analysis. By our modified P C R protocol he showed an abnormal pattern of heteroge-

neous repeats (ranging from 38 to 40) (Fig. 5). Following this observation we conducted a closer inspection of the Southern blot data and observed a faint band in the abnormal range indicating that approximately 5% to 10% of the cells have an enlargement of about 270 repeats that are methylated, thus identifying the patient to be mosaic. The patient had an affected brother whom we found to have an allele of approximately 300 repeats. This confirms that the mother is an obligate carrier. The patient's alleles in the normal size range could have arisen by deletion form the enlarged allele, and the heterogeneity in size could be related to that event or it could have arisen subsequently. Nevertheless, our modified protocol and detection system were able to detect this case, which was missed earlier.

Discussion P C R amplification of D N A sequence containing stretches of trinucleotide repeats has been problematic because of the secondary structure present in these templates [8]. To overcome this problem, most conventional P C R protocols for amplifying triplet repeats include reagents that facilitate D N A strand separation such as 7-deaza-2'-deoxyguanosine triphosphate and dimethyl sulfoxide; however, the presence of 7-deaza-2'-deoxyguanosine triphosphate reduces the detection of P C R products stained with ethidium bromide [17] and thus requires an additional detection method such as probe hybridization [16] or inclusion of ~_32p_

Automated Detection of Trinucleotide Repeats

•

Hamdan et al.

./=.

©

•

O~

m

-g c~ O

X ii

J

IX: I:I:

J

J

i

~E

E . ~ o°

o~o

>,Z

~

~ ~ ~-~;

.

8

~o6 [

c~

~D,,~ '~

~

'~

267

268

Molecular Diagnosis Vol. 2 No. 4 December 1997

labeled deoxynucleotide triphosphates in the reaction for detection by autoradiography [7]. Nevertheless, despite the use of D N A destabilizing reagents, "stuttering" is not entirely eliminated in these protocols. The sizing of fragments in the conventional protocol for F R A X is done by the operator, who has to decide on the basis of visual inspection which band amidst the "stutterers" to call as the major band. This band is then used to determine the number of repeats relative to an M13 sequencing ladder run alongside. This practice can lead to inaccurate sizing, difficulty in discerning alleles that differ by only one or two repeats, and difficulty in discriminating abnormal patterns such as those found in mosaics. The use of betaine in PCR protocols to amplify difficult templates has been described for amplifying triplet repeats and also for sequencing difficult templates. A recent report identified D N A polymerase pause sites as often being located near putative hairpins that contain the sequence P y - G - C surrounded by a GC-rich sequence [18]. Addition of betaine at a final concentration of 2 M eliminated virtually all the P y G - C pauses in template s used in the described study. As CGG repeats in F R A X are likely to form stable hairpin structures [8] and also contain the P y - G - C motif, it is thus not surprising that use of betaine would result in reduction of pauses by Taq D N A polymerase, allowing complete amplification of CGG repeats. Besides decreasing the secondary structure of D N A templates, inclusion of betaine in PCRs does not interfere with visualization of D N A in agarose gels stained with ethidium bromide. Thus, the success of the PCR can be verified immediately after the cycling is completed and does not require a waiting period of 1 to 2 days as in autoradiography. Here we demonstrate the accuracy and sensitivity of our automated detection system not only in discerning alleles but also in discriminating difficult patterns in F R A X PCR. The most difficult female patterns to distinguish are homozygotes and heterozygotes differing by one repeat [16]. As is evidenced in this study, both of these difficulties were overcome using our protocol. In samples from females, when only one allele is discerned, there is always the possibility that the second allele is fully expanded and not amplifiable or that the patient is homozygous for a normal F R A X allele. By duplexing the F R A X with A R PCR, it can be rapidly determined if a patient is homozygous for a

normal allele or has one fully expanded allele. Southern blot analysis is then used to confirm the presence of a second fully expanded allele. Although alleles greater than 100 repeats cannot be detected using this protocol, this does not pose a serious problem because we routinely complement the PCR protocol with Southern blot analysis, which detects fully expanded repeats and hypermethylation [5,19]. We have validated and are now implementing this modified PCR and detection protocol in our routine laboratory operations for the diagnosis of FRAX. Results are obtainable in less than 24 hours with minimal hands-on time. The use of radioisotopes for PCR of F R A X has been completely eliminated. Other nonradioactive, highresolution protocols for F R A X have been described [10,16]; however these protocols require use of nonradioactive probes to detect PCR products. By analysis of PCR products in thin, denaturing acrylamide gels coupled with automated laser fluorescence detection, products can be detected immediately. Low signal peaks, as present in amplification of mosaics and large alleles, can still be detected because of the minimal background in the detection system. We have successfully used a similar protocol for detection of triplet repeats in the myotonic dystrophy gene (data not shown), which contains the triplet repeats (CTG)n [20]. Thus, to date we have been able to detect three classes of repeats [(CGG)n, (CAG)n, (CTG)n] using this protocol. We believe our modified protocol can be used to amplify and detect most triplet repeats. We envision that a standard protocol for triplet repeats can be designed using a standard reagent mix for PCR in which only the amplification primers differ depending on the triplet repeats being amplified. These products can be detected in an automated gel sequencing system with the use of the appropriate molecular weight standards. Compared with the conventional protocol, this modified protocol is rapid and offers better resolution and direct detection of products without use of radioisotopes or hybridization probes. Several automated laser fluorescence sequencing systems are now available at a reasonable cost. Use of an automated detection system allows the automated calculation and storage of data that could be imported directly into a computerized reporting system. The monitoring of all high-risk pregnancies for F R A X has been pro-

Automated Detection of Trinucleotide Repeats

posed [16]. We believe that an a u t o m a t e d p r o t o c o l such as the one we have described h e r e will m a k e such m o n i t o r i n g m o r e feasible.

Acknowledgments We t h a n k A n t h o n y Bailey for excellent technical assistance, John Martinez for insightful discussions on the use of betaine, and R o n K a g a n for reading the manuscript.

Received August 5, 199Z Received in revised./brm August 2.5, 199Z Accepted August 27, 1997.

References I. Kremer E J, Pritchard M, Lynch M, Yu S, Holman K, Baker E, Warren ST, Schlessinger D, Sutherland GR, Richards RI: Mapping of DNA instability at the fragile X to a trinucleotide repeat sequence p(CCG),,. Science 1991;252:1711 1714 2. Bates G, Lehrach H: Trinucleotide repeat expansion and human genetic disease. Bioessays 1994;16: 277 284 3. Caskey CT, Pussuti A, Fu YH, Fenwick RG, Nelson DL:Triplet repeat mutations in human disease. Science 1992;256:784 789 4. Fisch GS, Nelson D L Snow K, Thibodeau SN, Chalifous M, Holdcn JJA: Reliability of diagnostic assessment of normal and prcmutation status in the fragile X syndrome using DNA testing. Am J Med Genet 1994;51:339 345 5. Hornstra IK, Nelson DL, Warren ST, Yang TP: High resolution methylation analysis of the FMR1 gcne trinucleotide repeat region in fragile X syndrome. Hum Mol Genet 1993;2:1659 1665 6. Brown WT, Jenkins EC: The fragile X syndrome. In Friedmann T: Molecular genetic medicine, vol 2. Academic Press, San Diego, 1992, pp. 39-66 7. Fu YH, Kuhl DPA, Pizzuti A, Pieretti M, Sutcliffe JS, Richards S, Verkerk AJMH, Holden JJA, Fenwick RG Jr, Warren ST, Oostra BA, Nelson DI, Caskey CT: Variation of the CGG repeat at the fragile X site results in genetic instability: resolution of the Sherman paradox. Cell 1991;67:1047 1058 8. Gacy AM, Goellner G, Juranic N, Macura S, McMurray CT: Trinucleotide repeats that expand in human disease form hairpin structures in vitro. Cell 1995;81:533-540

°

Hamdan et al.

269

9. Papp AC, Snyder P J, Sedra M, Guida M, Prior TW: Strategies for amplification of trinucleotide repeats: optimization of fragile X and androgen receptor PCR. Mol Diagn 1996;1:59-64 10. Baskaran N, Kandpal RR Bhargava AK, Glynn MW, Bale A, Weissman SM: Uniform amplification of a mixture of deoxyribonucleic acids with varying GC content. Genome Res 1996;6:633-638 11. Yancey PH, Clark ME, Hand SC, Bowlus RD, Somero GN: Living with water stress: evolution of osmolyte systems. Science 1982;217;1214-1222 12. Rees WA, Yager TD, Korte J, yon Hippel PH: Betaine can eliminate the base pair composition dependence of DNA melting. Biochemistry 1993; 32:137-144 13. Buffone G J, Darlington G J: Isolation of DNA from biological specimens without extraction with phenol. Clin Chem 1985;31:164-165 14. Maniatis T, Fritsch EF, Sambrook J: Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1982 15. Snow K, Tester D J, Kruckeberg KE, Schaid DJ, Thibodeau SN: Sequence analysis of the fragile X trinucleotide repeat: implications for the origin of the fragile X mutation. Hum Mol Genet 1994; 3:1543-1551 16. Brown WT, Houck GE, Jeziorowska A, Levinson FN, Ding X, Dobkin (7, Zhong N, Henderson J, Brooks SS, .lcnkins EC: Rapid fragile X carrier screening and prenalal diagnosis using a nonradioactive PCR lest. JAMA 1993;27/):1569 1575 17. Innis MA: PCR with 7-deaza-2'-deoxyguanosine triphosphate. In lnnis MA, Gclfand DH, Smisky JS, White TJ: PCR protocols: a guide to methods and applications. Academic Press, Orlando, FL, 1990, pp. 54 59 18. Mytelka DS, Chambcrlin MJ: Analysis and supprcssion of DNA polymerase pauses associated with a trinucleotide consensus. Nucleic Acids Res 1996; 24:2774-2781 19. Rousseau F, Heitz D, Biancalana V, Blumfeld S, Kretz C, Boue J, Tommerup N, Der Hagen CV, DeLozier-Blanchet C, Croquette MF, Gilgenkrantz S, Jalbert R Voelckel MA, Oberle 1, Mandel JL: Direct diagnosis by DNA analysis of the fragile X syndrome of mental retardation. N Engl J Med 1991;325:1673-1681 20. Fu YH, Pizutti A, Fenwick RG Jr, King J, Rajnarayan S, Dunne PW, Dubel J, Nasser GA, Ashizawa T, DcJong R Wieringa B, Korneluk R, Perryman MB, Epstein HE Caskey CT: An unstable triplet repeat in a gene related to myotonic muscular dystrophy. Science 1992;255:1256 1258